Approach 1 applies a neural network in Keras. The analysis builds on the example analysis to classify credit card fraud by Fancois Chollet (https://keras.io/examples/structured_data/imbalanced_classification/).
Approach 2 uses imbalanced classification methods implemented in sklearn based on https://towardsdatascience.com/machine-learning-and-class-imbalances-eacb296e776f
This is work in progress.
The data used here can be visually inspected at https://idheapunitepolsoc.shinyapps.io/covid_dashboard/